Every operating system includes a file system to manage data files. An API is provided by the operating system to develop applications providing an interface to the user to manage his own files. A typical application is a file manager to create, move, copy, delete, and rename files interactively through a user friendly graphical interface. Other applications such as text or graphical editors also use the API of the operating system to allow the user to manipulate the data he has processed.
The file system assigns to the users a hierarchical organization of their files. In such a tree data structure, the nodes are folders having branches with other folders, the leaves being the files. Each node and file is identified by a name which is a character string having a limited size and that the user is free to assign. The file or folder path is the list of the node names of the system file hierarchy to reach it, the last name being the file or folder name. With the applications using the file system API, the user can first visualize a hierarchical representation of his files. The graphical view is a tree in which the nodes are identified by the folder and file names. The user selects one folder or file by clicking on the name. The structure depending on one node can be either expanded or collapsed.
The applications using the file system API such as the file manager allow the user to create a file. At the file creation, the file manager requests the user to give a name. This name is unique for each file. To retrieve an existing file, the user can navigate, through the graphical representation of the file system hierarchy and select it by clicking on the corresponding node. Retrieving a file knowing its name is immediate. To retrieve a file without knowing its name requires knowledge of its path or at least the name of nodes belonging to its path to find it quickly in the file system hierarchy.
Usually, when a file is created, the user faces different possibilities to choose the node in the hierarchy where the file is to be stored. Most of the time, the file has a relationship with its content, the recipient, the product described in the file, and multiple other possibilities. At creation time, the recipient information may appear to be the most relevant information and the user chooses to store it in the file hierarchy containing everything related to recipients (customers for example). Unfortunately, today, one path must be selected and only one. At retrieval time, which may be months later, the relationship the user remembers might not be the recipient but instead the product described and therefore the user will start looking for the file in the hierarchy under products without a chance to find it.
Today, file managers reference a file by its unique combination of path and file name. With the U.S. patent Of Sun Microsystems, Inc., U.S. Pat. No. 6,055,540, a method is disclosed to create a class hierarchy. The patent proposes a document classification on top of the usual tree data structure built in the file managers, to perform a more efficient retrieval of documents. With the embodiment described, the user can create a class hierarchy by defining a tree structure of category nodes, the leaves and a category definition. The method provides also an interface to update the category node structure or definition. The solution of the prior art is helpful to be mapped on an application database such as a software engineering database used by application developers. It is used, for instance, to organize the access to data between developers as the categories of data are created in relation with development. This solution of the prior art requires time and process resources to manage the category nodes. It is used for coordinating a group of users and to organize the data as they are created and updated by this group of users.
The standalone user of a computer managing his own file system does not want to create and manipulate a new data structure over the standard tree data structure proposed by the file manager of the operating system. Neither does the user want to store the files according to a list of categories that need to be shared by a group and need to be managed as with the solution of the prior art.